Amazon seems to have quietly launched a market place for online / telehealth clinics. Didn't know this was coming, but seems like a pretty big deal!
clinic.amazon.com/?ref_=EM_ACQ_8 h/t
Follow
Click to Follow AndrewLBeam
Andrew Beam
@AndrewLBeam
Machine learning for medicine. Assistant Professor: , , || Co-host NEJM AI Grand Rounds (), Co-founder
Andrew Beam’s Tweets
Illuminating and fun (and at times high entropy!) conversation with Michael Abramoff and coming to Grand Rounds soon
1
3
18
We have extended the submission deadline to Feb 15th, 11:59pm EST.
Please share with your networks and visit chilconference.org for more details!
9
13
Quote Tweet
New preprint!
Q: How accurate is GPT-3 at predicting Dx and triage advice compared to Harvard docs and to an average person Googling their symptoms?
For Dx, GPT-3 is much more accurate than your average person and almost as accurate as the Harvard docs! For triage, not so much.
Show this thread
3
4
I found these results pretty interesting! The Dx performance was good, but it's worth thinking about what's up with triage, since that is arguably the best use case for something like GPT-3.
If you are interested, read the full paper for more details:
1
2
12
Show this thread
We used a small set of well-validated, synthetic patient vignettes confirmed to not be part of GPT-3's training data.
We compared GPT-3's performance to a sample of attending internal medicine docs and to a random sample of 5,000 people who were allowed to use the internet.
1
6
Show this thread
New preprint!
Q: How accurate is GPT-3 at predicting Dx and triage advice compared to Harvard docs and to an average person Googling their symptoms?
For Dx, GPT-3 is much more accurate than your average person and almost as accurate as the Harvard docs! For triage, not so much.
7
26
89
Show this thread
As a researcher in the healthcare field, I often find it tedious to keep track of conference deadlines. To solve this issue, We developed a website to easily track healthcare conf & workshops, integrated with Google Calendar for notifications.
deadlines.openlifescience.ai
#healthcare
2
8
16
Show this thread
📢 New Clinical LLM alert!
T5-Base & T5-Large models trained from ✨scratch✨ on MIMIC III + IV notes
Check out physionet for access: physionet.org/content/clinic
This Tweet is unavailable.
14
65
How much does a child's gestational age impact their cognitive abilities later in life? 🧠
Our study (bmj.com/content/380/bm) published in the BMJ () "Gestational age at birth and cognitive outcomes in adolescence" digs into this question.
Read 🧵 for our results👇
2
22
51
Show this thread
Looks like there is now a path to use the OpenAI tech stack (GPT-3, ChatGPT, etc) on data with PHI:
Quote Tweet
Healthcare AI builders
@Microsoft Azure @OpenAI Service is now HIPAA ready (+ includes support for zero-shot or specialized model creation via fine tuning on your own data)!
azure.microsoft.com/en-us/products
Check out this content for tips: lnkd.in/dUVZC6Ki
6
18
Quote Tweet
Here are 7 more reasons to grab your ticket to Delphi! We've got 7 more speakers joining the conference, and more to be announced!
March 27-29, 2023
Fort Lauderdale, Florida
buff.ly/3XJXXSC
#MedEd #EBNeo #NICU #Neonatologia #Neonatology #Neonatology
March 27-29, 2023
Fort Lauderdale, Florida
buff.ly/3XJXXSC
#MedEd #EBNeo #NICU #Neonatologia #Neonatology #Neonatology
8
It's story time!
Find me on the other side of the microphone as a guest on the Podcast, hosted by & .
I share lessons learned from my journey as a researcher in AI, and my bets for the future of medical AI.
2
15
39
Episode 2 of NEJM AI Grand Rounds is here!
and are joined with , an assistant professor who leads a lab focused on developing AI capable of highly complex medical decision making.
Listen now:
7
151
1,266
On the next episode of AI Grand Rounds, we sit down with the prodigious Pranav Rajpurkar () to talk about AI in radiology, what he learned about mentorship from , and the impact of large self-supervised models on medicine.
Link:
3
18
This year's #ClinicalNLP will be colocated with #ACL2023 in Toronto will host the MEDIQA-Chat shared task! Details below... twitter.com/AsmaBenAbacha/
This Tweet is unavailable.
2
21
I'm teaching an 8-lecture causality class for MIT's IAP term! The course website is here: github.com/csquires/6.S09.
The first two lectures are recorded and the rest will be as well. The first lecture is here: youtube.com/watch?v=tOguq_.
10
179
964
Show this thread
ChatGPT is rising rapidly through the academic ranks - from middle author to corresponding in just a few days!
sciencedirect.com/science/articl. h/t
Quote Tweet
Check out the author list h/t @AndrewLBeam
Performance of ChatGPT on USMLE: Potential for AI-Assisted Medical Education Using Large Language Models | medRxiv medrxiv.org/content/10.110
3
9
29
MIMIC-IV was published this week. The core dataset has been out for a while, but we've just published the deidentified free-text clinical notes: 300,000+ discharge summaries and 2.5 million radiology reports!
16
157
519
Show this thread
Sign up for our companion newsletter featuring content on how AI will change healthcare, its impact on the patient experience, and the people pushing for innovation:
5
7
Delighted to share our new paper at the intersection of LLMs + health.
Our LLMs building on Flan-PaLM reach SOTA on multiple medical question answering datasets including 67.6% on MedQA USMLE (+17% over prior work).
arxiv.org/abs/2212.13138
37
461
2,055
Show this thread
Who isn't represented in AI task design? Talented student (Michael Chen) looked at the diseases represented in #dermatology AI tasks in our paper:
2
6
26
Show this thread
Just had a lot of fun sitting down with Lily Peng of on AI Grand Rounds to talk about her path-breaking work on medical AI for ophthalmology!
1
26
This was a really fun conversation with two upcoming audio media stars (who also happen to be quite good at #AI)
twitter.com/NEJM_AI/status
This Tweet is unavailable.
2
1
19
How does machine learning accelerate 'omics and particularly genomics? Nice weekend listen Prof. on new
podcast spoti.fi/3PAidmf with hosts Profs.
7
15
Excited to release with my co-host the first episode of Grand Rounds -- an enlightening and fun conversation with of
1
2
16
The first episode of NEJM AI Grand Rounds is live! Listen as our first guest, Associate Dean , shares how his team is using #ArtificialIntelligence to prevent, predict, and beat disease:
9
23
The first episode of AI Grand Rounds is out!
We sat down with to talk about genomics, the future of AI in cardiology, and what he thinks aspiring clinicians should know about AI:
4
14
1
17
26
New preprint w/ where we proposed a new regularization scheme for CLIP-style models.
The method, named TIER, is applicable to most CLIP-style models and results in SOTA performance on a broad array of zero-shot CXR tasks!
Read more here: arxiv.org/abs/2212.06710
4
18
Meet the hosts:
, PhD, is an assistant professor of epidemiology at and the . The Beam lab develops new deep learning & causal inference methods for medical decision making, and they have a special interest in neonatal and perinatal medicine.
read image description
ALT
1
7
39
Show this thread
Meet the hosts of NEJM AI Grand Rounds:
, PhD, is an assistant professor of biomedical informatics . Raj directs a research lab of #machinelearning scientists, clinicians, and biomedical data scientists working to improve medical decision making.
read image description
ALT
2
7
30
Show this thread
NEJM AI Grand Rounds, a new podcast from NEJM Group exploring #ArtificialIntelligence and its application to medicine, is coming later this month. Listen to a preview and subscribe: ai-podcast.nejm.org #AIinMedicine
2
27
62
Our new machine learning model, Chroma, can generate extremely large and intricate proteins and #protein complexes in just a few minutes. Read the preprint on and learn about our breakthrough in protein #science: ow.ly/A0oQ50LYIPg #ProteinEngineering
1
31
150
study led by , Joe Hakim & Sonia Hernadez-Diaz (et al.) on effectiveness of 17-OH for prevention of recurrent preterm birth was cited as evidence by in their recommendation to remove preterm drug from market.
Learn more: causalab.sph.harvard.edu/causalab-news/
1
2
14
Full pre-print up on BioRxiv! Super proud of my colleagues and to be a part of the team revolutionizing medicines through generative biology. Illuminating protein space with a programmable generative model | bioRxiv
1
2
66
Come stop by our poster today to know about our work on using conformal prediction framework to identify reliable zero shot classification tasks in self-supervised models trained with contrastive loss.
Work done with and
2
14
Interested in conformal methods for self-supervised models such as CLIP? Come by and chat with and at today's workshop on self-supervised learning.
Workshop: sslneurips22.github.io
Paper:
1
2
12
Interested in biomedical LLMs or multi-modal learning for precision health? We Health Futures are too!
1
27
116